[whoisthis4chan]

> Typical gravitational-wave events change the length of the arms by only a fraction of the width of a proton. Sensing such minute changes requires painstaking isolation from noise coming from the environment and from the lasers themselves.

i find it utterly fascinating that we're able to detect such a minuscule deviation

[dekhn]

interferometry is indeed amazing. When the ultra-important Michelson-Morley experiment was run some ~100 years ago, they were doing interferometry but in those days there wasn't really good vibration isolation technology. They had to float their whole experiment on a pool of mercury (!) in the sub-sub basement of an idle building, and even then, deliveries nearby (by horse) would cause problems.

Nowadays, physics students do the MM experiment in a lab on a benchtop in a day.

[acqq]

I'd like to read how these problems are solved "in a lab on a benchtop" today!

[dekhn]

the original experiment is pictured here: https://en.wikipedia.org/wiki/Michelson%E2%80%93Morley_exper...

what makes it possible to do in a desktop lab course combination of a large number of different innovations. The first is that we know how to make extremely stiff/rigid/strong/flat/thermally stable tables (https://www.thorlabs.com/navigation.cfm?guide_id=41) which can optionally be placed on active vibration-cancelling struts (https://www.thorlabs.com/newgrouppage9.cfm?objectgroup_id=10...). The second is using cage systems for mounting things with everything lined up parallel and centered (https://www.thorlabs.com/navigation.cfm?guide_id=2255). The third is precise kinematic mounts which make real-time angle tuning a lot easier/more reliable (https://www.thorlabs.com/thorproduct.cfm?partnumber=KM100#ad...). The fourth is now we have powerful lasers and LEDs that make generating lots of light all pointing in the right direection easier (https://www.thorlabs.com/thorproduct.cfm?partnumber=CPS532-C...). The fifth is that high quality standardized optical parts (mirrors, lenses, etc) are easily available from a wide range of vendors (https://www.thorlabs.com/newgrouppage9.cfm?objectgroup_id=10...).

There are a number of other innovations in material science. but I'd recommend taking a look at Thorlab's Michelson-Morley educational kit. For $3K you get basically everything you need to carry out the experiment: https://www.thorlabs.com/thorproduct.cfm?partnumber=EDU-MINT... plus a nice manual https://www.thorlabs.com/drawings/5d9e11209b7d4536-820A3379-... that walks you through physical setup and theory behind the experiment (which among other things helped lead to special relativity).

if you want more like this, see https://www.thorlabs.com/newgrouppage9.cfm?objectgroup_id=11... which is a hardware kit that accompanies an actual optical lab class. The course is online: https://www.thorlabs.com/drawings/5d9e11209b7d4536-820A3379-... and gives a fairly straightforward introduction to optics. With this, you can easily build a microscope from components or any number of other nifty optical systems.

Non-optics people (IE, programmers, etc) with enough time and money can learn how to do real-world optical experiments in their garage (this applies to astronomy too). For example after a significant time/money investment, have started building my own microscopes which use real-time object detection to track tardigrades to do behavior analysis (lest anybody feel imposter syndrome, trust me it took a ton of time and money and even then I'm not quite at the level of a good grad student).

It's not my favorite but you can also read https://www.amazon.com/Perfectionists-Precision-Engineers-Cr...

If you want to truly go down the rabbit hole, https://pearl-hifi.com/06_Lit_Archive/15_Mfrs_Publications/M...

[acqq]

Wonderful answer, thanks!

Do you know if the "Michelson-Morley educational kit" is really enough to achieve the accuracy of the original experiment or is it just to make "any" functioning interferometer?

[dekhn]

I'm pretty sure it exceeds the accuracy of the original experiment. I think not being based on a trough of mercury is pretty important as well. But the manual shows several types of interferometers that can be built in lab courses.

[acqq]

Still, I see it is actually called "Michelson Interferometer Educational Kit", not "Michelson–Morley" and the user guide I'm reading (your link gives "The resource you are looking for has been removed", so I've clicked on the "User Guide" on the page instead) also takes care to never directly mention Morley or to suggest that the same experiment can be reproduced with that kit.

[fsh]

There is no way you could repeat the Michelson-Morley experiment with that small and floppy Thorlabs EDU kit. The experiment from 1887 had an arm length of 11 m and was interferometrically stable (typical length fluctuations much smaller than the wavelength) while rotating. That would still be a considerable engineering challenge today.

Modern Michelson-Morley experiments [1, 2] don't use Michelson interferometers anymore. Instead, they compare the lengths of crossed ultrastable high-finesse cavities (in vacuum, of course). The big innovation is that, with lasers and electronics, we can measure the cavity resonance frequencies (and therefore also the cavity lengths) to something like 15 digits of accuracy. This corresponds to less than a tenth of the diameter of a Proton, and is something like 100 million times more accurate than you can achieve with a simple Michelson interferometer.

[1] https://doi.org/10.1103/PhysRevLett.103.090401

[2] https://doi.org/10.1103/PhysRevD.80.105011

[dekhn]

Thanks. I'm not an optics expert, although my friend (the one who said they built a michelson interferometer in a day in their physics lab) is. Since I don't want to mislead anybody, could you explain what the Thorlabs kit can do, and is it technically a Michelson interferometer? The labs that feature the kit all seem to measure the wavelength of light.

It would seem odd that Thorlabs (generally well respected) would sell something that is not what it really is, or misrepresented its capabilities. my guess is that you're sayting the kit itself couldn't reproduce the original experiments, but that it still is a Michelson interferometer in design, which can be used to carry out less demanding experiments, but not demonstrate the (non) existence of aether?

[fsh]

The Thorlabs kit looks like a very decent Michelson interferometer that can be used for a lot of demonstrations such as measuring wavelengths and studying the coherence properties of light sources.

However, repeating the Michelson-Morley experiment is not easy since the expected signal is very small. If there was a stationary aether, the relative length difference for the optical path along the earth's motion compared to the path perpendicular would be (v/c)^2 ≈ 1E-8, where v is the orbital velocity of earth (3E4 m/s), and c is the speed of light (3E8 m/s). The arm length of the Thorlabs kit is just a few cm, so the shift would be on the order of one nm, or one five-hundreth of a (green) wavelength. Thermal drifts and vibrations of optics on a typical optical table are much larger than that, especially when trying to rotate the setup. Michelson and Morley achieved the necessary stability by constructing their interferometer on a solid stone slab, and made a near-frictionless bearing by floating it on mercury. The resulting stability is still impressive by modern standards, but the construction technique is not very practical. Nowadays, large and passively stable optics setups (for example telescope mirrors or laser gyros) are usually made from massive pieces of Zerodur which has near-zero thermal expansion.

[acqq]

"the one who said they built a Michelson interferometer in a day in their physics lab"

Maybe you should check with them too: it's possible that they have also built "a" Michelson interferometer (just like Thorlabs kit features one) but maybe their setup was in spite of that insufficient to perform the needed measurements in the way needed for the valid execution of a Michelson-Morley experiment?

Historically, Michelson constructed his first interferometer in 1881 in Potsdam, Germany:

https://arxiv.org/abs/2111.12176

Inventing it was obviously necessary but not sufficient for a valid Michelson-Morley experiment, which was correctly finished only during 1887 in Cleveland, Ohio.

[moh_maya]

/off topic

Have you documented & written about your microscope experiments (both building them and the experiments themselves) on the web?

If you have, and are comfortable, can you please share the link?

Just curious. Ex-biologist, and among other things, I helped build / assemble an optical trap and other equipment including confocal / TIRF microscopes about a decade+ ago, so curious what a serious (amateur) student with time, passion and resources is able to do.

Thank you :)

[dekhn]

No, I don't document and write about my experiments. I've shared them with a tardigrade biologist, but we didn't decide to continue. Also, I'm not a student, I'm a professional computer guy with a hobby and a budget. That said, smart undergrads coudl definitely do this.

Nothing I'm doing is remarkable or complicated, and compared to a research microscope, what I've done is very trivial. It's just a 10X scope with a grbl-controlled XYZ stage and an object detector that finds the center of a tardigrade in realtime, and sends commands to center the tardigrade. Now my interesting is in high speed scanning- instead of taking photos, you literally take a video while actively moving the head around and then stitching it all together. Not stopping to take pictures increases the rate of acquistion 10X or more.

The reality is that I could have done everything I wanted to do by spending about $10K for a kit from Thorlabs, but I was interested in learning more about building precision stages from inexpensive components, so that if/when I ever do get to play with the expensive toys, I know why they are better, anbd why they cost so much money.

The interesting area now is SPIM see https://openspim.org/Step_by_step_assembly which is definitely something a well-funded hobbyist could do.

[moh_maya]

Thank you!

[funac]

you can build very good hydrostatic vibration isolators in a home machine shop nowadays; commerical optical tables are /very/ steady

[pfdietz]

Lasers beams are bounced back and forth many times, so the deviation builds up. The beams have to be very powerful (100s of kW) to reduce photon counting noise sufficiently.

[wwarner]

Kip Thorne explains it pretty clearly in this 2002 lecture

https://youtu.be/mGdbI24FvXQ